[Crm-sig] non-existent objects
maximilian at schich.info
Wed Nov 7 18:29:52 EET 2012
Pre-Cidoc-CRM, this problem is traditionally solved with the following
1. "unspecified" (if the field is still empty, meaning nobody thought
2. "none" (if non-existence is a proven fact)
3. "unknown" (if existence is possible)
The values are instances, not part of the model or namespace, so
database curators can use them no matter how you define the data model
in the first place.
Note: While "none" may have one URI, "unknown" should be disambiguated,
as not all unknown concepts or statues are the same (and archaeologists
sometimes find two consecutive setups of the same unknown statue).
Also note: Yes, you should expect a two digit percentage of "unknown"
URIs in archaeological datasets. By sheer number this makes them
instances, not part of the model.
Three critical observations:
* "sheep" has a plural morpheme: It is "sheep".
* "without base" is just as valid a concept as "base". Kids often say
"pasta with nothing". A concept with an URI does not necessary need
a physical equivalent.
* Personally, I also prefer Wolfgang's first model over the second.
But in general, model configuration on this level can't be
standardized. There will always be differences in approach by
different scholars. For e.g. Simpson's, Jantzen's, and Sedlmayr's
account on Gothic Cathedrals are essentially different due to such
differences in perspective. To standardize them into one framework
would simply miss the point.
On 2012-11-07 16:47 , Athanasios Velios wrote:
> Again a modest and relatively uninformed contribution on my behalf:
> We have lots of similar issues in the description of bookbinding
> structures, where features are missing and the problem is more
> complicated than that.
> What you describe could be considered as the condition of the statue:
> e.g. missing. You are certain that the statue did exist. As we often do
> in conservation, you assessed the statue's condition and concluded that
> the whole statue is missing. So the uri should be valid because the
> statue existed (and may well still exist).
> The difficulty starts when people want to describe a base for which a
> statue never existed, say a half-finished base or a test base. The fact
> that a statue never existed is now unrelated to condition yet equally
> significant. I think the solution to this problem is to create a new
> subclass of "base", something like "bases without statues" in your list.
> I hope the above is relevant.
> All the best,
> On 07/11/12 15:17, Wolfgang Schmidle wrote:
>> Dear all,
>> I am working on Arachne's Cidoc representation, and we came across a
>> problem with non-existent objects and how to state their non-existence.
>> A statue may be set up using e.g. a base or a plinth. In Arachne this
>> can be specified in a data field called "Aufstellung" ("setup"). One can
>> choose a description from a fixed list, for example "Basisplatte" or
>> "Fußplatte/Plinthe". Now, we could model it as
>> E22 (the statue, without setup) P46i forms part of E22 (the statue
>> plus the setup) P2 has type E55 Type e.g. "Basisplatte"
>> but I am told that the setup should be seen as a part of the statue.
>> Consequently we are modelling it as
>> E22 (the statue, including the setup) P46 is composed of E22 (the
>> setup) P2 has type E55 Type e.g. "Basisplatte"
>> However, Aufstellung may also have the value "ohne Basis" ("without
>> base"). In this case the second E22 would denote a non-existent object,
>> and its Type "ohne Basis" would state the non-existence of this object.
>> (If the data field is left empty, we make no statement at all about the
>> Is this the right way to model it? And is there a problem in RDF with an
>> URI for a non-existent object?
>> Additional question: Does Cidoc have an opinion about the the exact
>> meaning of E22 P46 E22 P2 E55 "ohne Basis"? Let's take the word "sheep"
>> as an example, where the singular and plural forms are the same: one
>> sheep, two sheep. Is it comparable to A) "while most words have a plural
>> morpheme, the particular word sheep has none", or B) "for systematic
>> reasons we assume that all words have a plural morpheme, but for the
>> particular word sheep it is null"?
>> Crm-sig mailing list
>> Crm-sig at ics.forth.gr
> Crm-sig mailing list
> Crm-sig at ics.forth.gr
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Crm-sig